Lots of things are acceptable, but in my case this would not be optimal.
About 99% of the time I bring in a sketch, drawing, picture, etc. and place it on one of the principal planes. The image is then offset so that the (for example) x-y position is where I want it, and the (for example) z position is slightly below the plane. Then I create a sketch on the same plane as the image.
I can then "trace" over the image to get a good start on the sketch without endlessly trying to refer to an external set of coordinates, lengths, etc.
This simplifies getting the sketch into a close resemblance to the final requirement.
Therefore I like the idea of presenting an orientation selection feature to the user as a very early step in the import.
I would never use the current camera position.
Of course others may have different needs.
Gene